Obesity affects over 34 million adult Americans and increases the risk of cardiovascular disease, hypertension, diabetes, and some cancers. Despite the known health risks of obesity, success in its treatment has not been good. Caloric restriction has been effective in producing but not in maintaining weight loss. There is a great deal of controversy regarding the use of exercise in treating and preventing obesity. Much of the controversy arises because the benefits of exercise can only be seen over long periods of time and it is difficult to conduct long-term studies where the amount of exercise provided is accurately determined and where the effects of exercise on energy balance are accurately measured. There are no data available in the literature which show the effects of a known amount of exercise delivered for a long period of time on body weight and body composition. Rather, existing studies are limited by the short nature and the inability to verify the amount of exercise performed. The first aim of this study is to determine the effects of a supervised 18 month exercise protocol on body weight and body composition of moderately overweight college students. We hypothesize that exercisers will show lower body weights, lower body fat, and similar or greater amounts of fat- free mass than sedentary controls. The second aim is to determine the manner in which exercise alters energy balance, leading to the observed changes in body weight and composition. For energy intake, we hypothesize that exercisers will increase energy intake less than the increase in energy expenditure and will increase carbohydrate intake. We expect no increase or a small increase for energy intake and no change in carbohydrate intake for sedentary controls. For energy expenditure, we hypothesize that exercisers will increase total energy expenditure less than the increased cost of exercise, yet energy expenditure will be greater than sedentary controls. We hypothesize that exercisers will increase the ratio of fat/carbohydrate oxidation compared to controls. Lastly, we hypothesize that energy expended in spontaneous activities will remain constant or decrease slightly for exercisers, but if a decrease is shown, the decrease will not equal the energy expended in exercise. Spontaneous activity will remain constant or decrease for sedentary controls.